The relative accuracy of simultaneous vs. consecutive interpreting with respect to specific difficulties (false starts, incomplete sentences, unimportant modifiers, elements without direct 'equivalents' in the target language, digressions) predicted by the Effort Models was partially assessed by a comparison of simultaneous and consecutive renderings of the same English speech into French by 10 interpreters in each mode. As predicted, consecutive was found superior in incomplete sentences, and simultaneous was superior as regards digressions and unimportant modifiers. With respect to overall accuracy, simultaneous was clearly superior to consecutive in this study, but methodological considerations call for caution when generalizing.